文档介绍：山西财经大学
硕士学位论文
基于马田系统对金融危机预警指标的选择与研究
姓名:田菁
申请学位级别:硕士
专业:金融学
指导教师:沈沛龙
2011-03-15
基于马田系统对金融危机预警指标的选择与研究
摘要
近年来,金融危机频发,其产生的严重后果对于整个世界经济都产生了深
远的影响。在过去,由于美国次贷危机引发的全球性的金融危机不仅对金融领
域产生了严重的影响,对实体经济的影响也异常深远。基于金融危机带来的严
重后果与深远影响,人们试图运用模型对金融危机进行预测,最早始于
()和(),他们建立的模型在使用样本内数
据的条件下取得了良好的预测效果,但在预测之初却需要进行一系列的假设,
导致模型对于金融危机预测欠缺普遍适用性。在之后的金融危机预警模型的研
究中取得了一些成果,但在对危机进行预测的过程中往往使用多个指标,导致
预测的过程复杂,收集数据的工作也较为繁重,最终取得的预测效果也不尽如
人意。
因此,在经历了次贷危机的严重影响后,对于金融危机的研究再一次成为
各国经济学家关注的热点。金融危机预警模型作为金融危机预测的主要方法,
在构造模型的过程中变量的选择是构造预警模型的关键。因此,选择合理的经
济变量作为构造预警模型的指标对于金融危机的有效预测具有重要的现实意
义。本文以个国家数据作为样本,选取了代表宏观经济不同侧面的个经
济指标,利用多元变量系统诊断技术马田模型进行分析,在不降低预警模型预
测水平的前提下,有效地减少了预测指标的数量,这对提高金融危机预警模型
的预测水平和减轻前期的数据收集工作具有重要意义。

【关键词】金融危机预警指标马田系统
1
山西财经大学硕士学位论文
ABSTRACT
In recent years, the financial takes place frequently, and its consequences for the whole world
economy has exerted profound influence. In the past, because of the . subprime mortgage crisis
triggered global financial crisis not only affects the financial sector, the impact to economic entity
also abnormal severity. Based on financial crisis brought serious consequences and far-reaching
influence, people tried to use models to predict the financial crisis began in the earliest Blanco and
Garber (1986) and Edin and Vredin (1993), they build model in the use of in-sample data obtained
good prediction effect, but it needs a series of assumptions, make model does not have to predict the
universal applicability of financial crisis. Next to the financial crisis early-warning system research
has acquired some achievements, but the crisis of predicting process often to use several index,
which causes prediction of the process plicated and the work of collecting data is also heavy,
the final prediction effect is not satisfactory
Therefore, in today of the subprime crisis has not been disappeared , for financial crisis
research again e the focus in all countries economists. Financial crisis early-warning